An Identity-Oriented Framework for AI-Supported mHealth

Metabolic chronic conditions require sustained lifestyle change, yet mobile health (mHealth) interventions typically show modest and short-lived effects. A central unresolved question is how behavior change becomes personally meaningful. Beyond meaningful gamification, a growing body of literature emphasizes identity processes as crucial for long-term change: health behaviors persist not merely through motivation or prompts, but when narratively integrated into the self-concept. In the age of large language models (LLM)-based conversational systems, this perspective becomes urgent. While LLMs can scaffold reflection at scale, it may normatively script users’ developmental trajectories. Existing work has addressed related questions around identity, reflection, and behavior change, but an integrative framework for identity-oriented mHealth design remains underdeveloped. This paper proposes an identity-oriented framework for designing mHealth systems that support narrative identity development and sustained behavior change. The paper organizes empirically reported design patterns along four narrative identity processes (interpretation, temporal integration, episodic configuration, and self-representation) to offer a design-relevant conceptual lens for LLM-supported mHealth systems.